We rounded up a few articles you may find interesting in this edition of Neal News. Check out the latest from our team on Reinforcement Learning, AI workloads on the edge, and data migration.

What makes a successful reward function?

Deep Reinforcement Learning (DRL) solutions rely on reward functions to train AI agents. Our Associate Practice Director of Data Science & Applied AI, Jayson Stemmler, offers a few tips for creating good reward functions in this blog post.

How to create RL-ready simulators

Simulators play an important part in RL solutions and getting the setup right can save your team a fair bit of time and effort. Check out this blog for the 4 key factors of RL-ready simulators, as well as some common pitfalls to avoid.

5 things for retailers to consider before deploying vision AI at the edge

Technologies like Azure Percept are making vision AI at the edge more accessible across industries. Retailers can use these solutions to monitor stock and in-store foot traffic, create customer arrival alerts, and improve supply chains. For a smooth deployment, make sure to consider these 5 things in your strategy.

3 reasons to migrate on-premises Spark workloads running on Hadoop to Azure Databricks

Why migrate these workloads from Hadoop to Databricks? Cloud benefits can help businesses resolve common on-prem challenges around data gravity, workflow inefficiencies, and data consistency. Here are our top 3 reasons to migrate.

Customer story: Designing and implementing a seamless ERP migration to the cloud

The Ciruli Brothers knew they needed a solution. With critical ERP data in an on-premises server, power outages and system failures could cripple operations for the entire company. Here’s how we helped the Ciruli Brothers rapidly migrate to the cloud, providing improved security and accessibility with Azure.

In other news…

• Wondering how digital twins, reinforcement learning, and simulators can improve product delivery? Check out this example from AnyLogic.
• Azure Percept helps you quickly deploy AI workloads to edge devices. See how it works in this video with Azure Lead Engineer George Moore.
• If you’re weighing the pros and cons of migrating Hadoop to the cloud, this Databricks whitepaper may help: The Hidden Value of Hadoop Migration
• Microsoft Build (May 25-27) registration is open! You can sign up for the digital event here.